Commit Graph

610 Commits

Author SHA1 Message Date
Evan Daykin
4f0cc20e2b Feature: add SIMM=RAW to ao records 2021-03-12 16:31:10 -05:00
Michael Davidsaver
504191441d Merge remote-tracking branch 'github/pr/99' into 7.0
* github/pr/99:
  fix an out-dated comment in the array filter code
  add macro dbfl_pfield to db_field_log.h and use it in dbGet
  fix in dbGet: decide use of db_field_log based on whether it has copy or not
  add macro dbfl_has_copy to db_field_log.h and use it in dbAccess.c
  refactor db_field_log and filters to get rid of dbfl_type_rec
  make it clearer what the result of wrapArrayIndices will be
2021-03-07 21:17:21 -08:00
Andrew Johnson
bbb4d86f78 Enable RTEMS testing in modules/database/test/std/link 2021-03-03 18:17:31 -06:00
Brendan Chandler
f41276bef8 epicPosicMutexInit: avoid calling with 0 which is platform dependent
Different platforms (RTEMS5) can define different values for
PTHREAD_MUTEX_DEFAULT, so we shouldn't pass 0 assuming its
PTHREAD_MUTEX_DEFAULT.
2021-03-02 22:44:56 -08:00
Andrew Johnson
f9e3e86401 Support VxWorks 6.9.x before taskWait() was added
We don't know exactly which version this was added in,
but it is present in 6.9.4.1 so use that.

Fixes lp: #1913699
2021-03-02 11:54:17 -06:00
Michael Davidsaver
3c46542630 posix: epicsMutexOsdShowAll check for PI support 2021-03-02 06:47:43 -08:00
Andrew Johnson
f8eb0be7a4 Update submodules after release 2021-02-28 21:39:28 -06:00
Andrew Johnson
a9ff175cf0 Replace "UNRELEASED" in epicsString.h 2021-02-28 17:37:40 -06:00
Andrew Johnson
458c0af4e9 Checkout submodules at tagged versions for 7.0.5 2021-02-28 17:36:25 -06:00
Andrew Johnson
eacee1b548 Merge Krisztian Loki's issue_97 branch into 7.0 2021-02-28 15:40:11 -06:00
Andrew Johnson
6754404d0f Replace magic "2" with macro AAI_DEVINIT_PASS1 2021-02-28 15:02:27 -06:00
Andrew Johnson
6734918e6e Release notes and aai documentation updates 2021-02-27 22:19:48 -06:00
Andrew Johnson
1c566e2110 Modify aai to support pass-1 device initialization
The Soft Channel device support requests pass-1 initialization.
It no longer needs to initialize the INP link or allocate the
array buffer itself, these are taken care of elsewhere.

The record code uses PACT to remember that the device must be
initialized again in pass 1.
2021-02-27 22:08:50 -06:00
Andrew Johnson
7f878d3074 Merge fix-1908305 branch into 7.0 2021-02-27 17:20:30 -06:00
Andrew Johnson
12ab69402a Improve the message from a DBD field-count mismatch 2021-02-27 17:15:47 -06:00
Andrew Johnson
08eaea64d2 Fix up comments & messages, add Release Notes 2021-02-27 16:48:55 -06:00
Andrew Johnson
bee00658ae Limit auto-declaration of record types to regRecDevDrv only
Allowing this while expanding DBD files for IOCs can insert other
device supports before of the Base "Soft Channel" ones, making the
other type the default. Adds a note that DBD file order matters.

Fixes lp: #1908305
2021-02-27 16:48:55 -06:00
Michael Davidsaver
e881cb15c4 registerAllRecordDeviceDrivers() handle function 2021-02-26 07:42:17 -08:00
Krisztián Löki
4a0f488657 Fixed db_post_events to not use bptr 2021-02-25 16:13:48 +01:00
Krisztián Löki
2340c6e6c1 Allow changing the BPTR field
This fixes github issue #97:
Reading into an aai record from a compress or histogram or subArray record
could cause a segfault if the aai record was initialized before the
linked one.
2021-02-25 14:54:07 +01:00
Michael Davidsaver
5593103c11 posix: epicsMutexShowAll() print if PI mutex are used
Confirms that pthread_mutex_t are actually being created
with PTHREAD_PRIO_INHERIT, ie. if both libc and kernel
support is present.
2021-02-21 20:01:45 -08:00
5a8b6e4111 posix: PI for epicsMutex, epicsEvent etc.
Enable priority-inheritance for primitives based on pthread_mutex_t.

Based on work by Till Straumann <till.straumann@psi.ch>
2021-02-21 20:01:45 -08:00
Andrew Johnson
9d0597fc15 Fix MS compiler issue that I introduced into Freddie's code
Plus a little more code simplification.
2021-02-21 12:04:07 -06:00
Andrew Johnson
93208af61c Merged Freddie Akeroyd's fix_win32_monotonic_time branch 2021-02-20 17:36:21 -06:00
Michael Davidsaver
552b2d1766 dbConstAddLink bounds check dbrType
cf. 6c914d19c3
2021-02-19 15:53:32 -08:00
Andrew Johnson
36d0fbd7be Another c89 compat fix 2021-02-08 13:34:01 -06:00
Michael Davidsaver
30e1431fb4 c89 compat 2021-02-08 09:43:51 -08:00
Michael Davidsaver
2c1c35268e db: Suggest DBF_MENU values on parser error 2021-02-08 09:22:27 -08:00
Michael Davidsaver
745c3f552e Com: add epicsStrSimilarity() 2021-02-08 09:22:27 -08:00
Michael Davidsaver
27918cb7a1 improve error message from dbPutString() for DBF_MENU/DEVICE 2021-02-08 09:22:27 -08:00
Michael Davidsaver
8723d4d9cb include database S_* in error string table 2021-02-08 09:22:27 -08:00
Michael Ritzert
410921b5ef Use getifaddrs instead of SIOCGIFCONF.
The old code is preserved in osdNetIfConf.c, which is used by the default
platform directory.

Platforms that support getifaddrs use the new code from osdNetIfAddrs.c.
2021-02-06 08:27:21 -08:00
Michael Davidsaver
43bd5ee1c2 Com: always skip SOCK_CLOEXEC for RTEMS and vxWorks 2021-02-05 10:32:47 -08:00
Andrew Johnson
78d685688c Add VALID_BUILDS support for "Command" 2021-02-02 16:06:42 -06:00
Michael Davidsaver
cf3173b6f4 posix: use SOCK_CLOEXEC and accept4()
If available, ensure O_CLOEXEC is set atomically.
Continue to F_SETFD as well (paranoia).

Available at least on Linux, freebsd, and RTEMS 5 w/ libbsd
2021-01-29 08:45:44 -08:00
Michael Davidsaver
1e471832e9 consolidate osiFileName.h
Don't really need 10 files for this
2021-01-27 09:19:13 -08:00
Michael Davidsaver
c19605232a epicsAtomicDefault.h typo 2021-01-18 08:23:54 -08:00
Ben Franksen
236bb2c671 fix an out-dated comment in the array filter code 2021-01-15 15:13:47 +01:00
Ben Franksen
372e937717 add macro dbfl_pfield to db_field_log.h and use it in dbGet 2021-01-14 17:45:25 +01:00
Ben Franksen
56f05d722d fix in dbGet: decide use of db_field_log based on whether it has copy or not 2021-01-14 17:40:37 +01:00
Ben Franksen
85822f3051 add macro dbfl_has_copy to db_field_log.h and use it in dbAccess.c
It encapsulates the slightly tricky logic to decide whether a pointer
to a db_field_log has ownership of the data or not.
2021-01-12 17:49:10 +01:00
Ben Franksen
27fe3e4468 refactor db_field_log and filters to get rid of dbfl_type_rec
This refactor simplifies and streamlines the code associated with server
side filters. Apart from immediate benefits (clearer code, less duplication)
it is also hoped that this will make it easier to add write filters.

The data pointer dbfl_ref.field can now either point to a copy owned by a
filter, or it can point to the original data owned by a record. In the
latter case, the dbfl_ref.dtor is NULL.

The dbExtractArray* functions are unified to the single function
dbExtractArray and stripped of conversion functionality. This is redundant
because we always call dbGet after applying filters, which takes care of
conversion. Accordingly, dbChannelMakeArrayCopy is now obsolete and its
single use (in the ts filter) replaced with dbExtractArray. Instead, we add
the helper function dbChannelGetArrayInfo to wrap the common boilerplate
around calls to the get_array_info method, used in both arr.c and ts.c.
2021-01-12 17:49:10 +01:00
Ben Franksen
4ab9808180 make it clearer what the result of wrapArrayIndices will be 2021-01-12 15:56:32 +01:00
Andrew Johnson
4e81eaa7e8 Update pvAccess module (caProvider) 2021-01-05 21:42:59 -06:00
Andrew Johnson
99852c6504 Merge JSON5 support into 7.0 2021-01-05 21:39:15 -06:00
Michael Ritzert
a9e3fa74aa FTBFS on FreeBSD 12 2021-01-05 08:56:57 -08:00
Andrew Johnson
f775e0b8f6 Merge 3.15 branch into 7.0
Adjust updated test result displays for submodules
2020-12-23 16:18:21 -06:00
Andrew Johnson
5f0d52cd5c errSymLib Fix from Ivo Hanak
Fixes lp: #1907077
2020-12-08 10:10:03 -06:00
Michael Davidsaver
4e055610b7 epicsUnitTest: SetErrorMode() 2020-12-03 08:33:57 -08:00
Michael Davidsaver
17ee7dd6c7 update PVA modules 2020-12-02 10:44:21 -08:00